• Laser & Optoelectronics Progress
  • Vol. 60, Issue 12, 1210019 (2023)
Qilong Yang1,2, Xiaoyu Ma2,3,*, Shuang Liu1,**, Shuanghui You2, and Chengping Li2
Author Affiliations
  • 1School of Optoelectronic Science and Engineering, University of Electronic Science and Technology of China, Chengdu 611731, Sichuan, China
  • 2Chongqing Co -Core Optoelectronics Technology Research Institute Co., Ltd., Chongqing400021, China
  • 3Chengdu Research Institute, Sichuan University of Arts and Sciences, Chengdu 635000, Sichuan, China
  • show less
    DOI: 10.3788/LOP221228 Cite this Article Set citation alerts
    Qilong Yang, Xiaoyu Ma, Shuang Liu, Shuanghui You, Chengping Li. Template Drift Suppression Method Based on Sub Pixel Correction[J]. Laser & Optoelectronics Progress, 2023, 60(12): 1210019 Copy Citation Text show less
    Sub pixel error in target tracking. (a) Original template position; (b) template position after offset and original template position
    Fig. 1. Sub pixel error in target tracking. (a) Original template position; (b) template position after offset and original template position
    Relationship between template drift and update times
    Fig. 2. Relationship between template drift and update times
    Schematic diagram of bilinear interpolation
    Fig. 3. Schematic diagram of bilinear interpolation
    Experimental scene and image. (a) Experimental scene; (b) experimental image
    Fig. 4. Experimental scene and image. (a) Experimental scene; (b) experimental image
    Comparison of template drift caused by different updating methods (MOSSE)
    Fig. 5. Comparison of template drift caused by different updating methods (MOSSE)
    Comparison of template drift caused by different updating methods (CSK)
    Fig. 6. Comparison of template drift caused by different updating methods (CSK)
    Comparison of template drift caused by different updating methods (KCF)
    Fig. 7. Comparison of template drift caused by different updating methods (KCF)
    Precision curves of different methods with default learn rate
    Fig. 8. Precision curves of different methods with default learn rate
    Success rate curves of different methods with default learn rate
    Fig. 9. Success rate curves of different methods with default learn rate
    Precision curves of different methods with large learn rate
    Fig. 10. Precision curves of different methods with large learn rate
    Success rate curves of different methods with large learn rate
    Fig. 11. Success rate curves of different methods with large learn rate
    Tracking results of six algorithms for different video sequences. (a) Dog1; (b) mhyang; (c) mountainbike; (d) sylvester; (e) trellis; (f) walking
    Fig. 12. Tracking results of six algorithms for different video sequences. (a) Dog1; (b) mhyang; (c) mountainbike; (d) sylvester; (e) trellis; (f) walking
    Video sequencefDSST /(frame·s-1fDSST_Cor/(frame·s-1Performance lost /%KCF/(frame·s-1KCF_Cor/(frame·s-1Performance lost /%CSK/(frame·s-1CSK_Cor/(frame·s-1Performance lost /%
    Average111.387.622.84191.2182.64.45329.5271.617.13
    dog1141.4125.011.60300.6289.73.63505.0419.616.91
    fish98.283.015.48128.3122.54.52332.9265.720.19
    mhyang109.590.317.53139.2133.14.38308.8250.019.04
    mountainbike97.165.033.06175.2167.24.57305.1245.719.47
    sylvester131.9109.816.76195.5185.65.06226.2204.79.50
    trellis72.442.741.0285.582.23.86198.9167.715.69
    walking128.697.124.45314.1297.95.16429.7347.719.10
    Table 1. Tracking speed of six algorithms for different video sequences
    Qilong Yang, Xiaoyu Ma, Shuang Liu, Shuanghui You, Chengping Li. Template Drift Suppression Method Based on Sub Pixel Correction[J]. Laser & Optoelectronics Progress, 2023, 60(12): 1210019
    Download Citation